Metrological aspects of collaborative field trials, including coping with unexpected events

International audienceWith the support of the French River Basin Agencies (RBAs), and the French Ministry of Ecology and its local offices, the French National Institute for Industrial Environment and Risks organized a collaborative field trial in June 2007 on a minor tributary of the River Seine near Paris. After sampling errors were recognized in the context of ISO 17025 as an important factor affecting the quality of an analytical result, the needs for consistency of data arising from the European Water Framework Directive 2000/60/EC rendered enhancement of metrological knowledge in this step of the measurement chain more significant. This trial was the first national attempt to improve knowledge of the effect in natural river waters of sampling activity undertaken as part of regulatory monitoring. Of the 14 sampling teams that took part in the trial, all selected by the RBAs, three were not linked to an analytical team. The Epone sampling spot was selected because of its central location, because its rural characteristics are representative of the majority of the RBA sampling stations, and because it enabled more than 40 people to be brought together in a secure environment. The monitoring parameters selected were : pH, oxygen, temperature and conductivity (field parameters); and, suspended particulate matter (SPM), nitrites, orthophosphates, barium, di(ethylhexyl) phthalate (DEHP), and one organic halogenated volatile (OHV) substance, tetrachloroethylene (parameters to be analyzed in a laboratory, also called non-field parameters). The aim of this study was to evaluate several sampling procedures, including standardized, to determine the variability induced by sampling operations in subsequent analytical processes. "Sampling activities" included sampling performance and external inputs (e.g., weather). individual field observations were also performed in order to understand the reasons for reference methods not being rigorously followed and to propose improvements that will make these methods more acceptable to sampling staff, and thus more reliable. This article focuses on the results obtained for non-field parameters and on the solutions that we developed to overcome the difficulties caused by a storm the day before the trial